15:02:07 #startmeeting fedora-qadevel 15:02:07 Meeting started Mon Jan 25 15:02:07 2016 UTC. The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:02:07 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:02:07 The meeting name has been set to 'fedora-qadevel' 15:02:07 #topic Roll Call 15:02:15 * mkrizek is here 15:02:31 * jskladan is here 15:04:00 * kparal is here 15:04:06 tflink: thanks for ping 15:04:41 np 15:04:46 let's get this party started 15:04:54 #topic Announcements and Information 15:05:05 #info Taskotron staging is running disposable clients - tflink 15:05:05 #info fmn.rules for taskotron fedmsgs - mkrizek 15:05:26 * kparal is still struggling with dnf metadata 15:05:48 the last image I built had the dnf makecache in $post 15:05:51 %post 15:06:11 great 15:06:24 probably should build a new one today 15:06:36 and start figuring out where to build them not on my machines :) 15:06:39 #info Pull request to fix ImgFac's raw image download url endpoint on Github 15:06:40 #link https://github.com/redhat-imaging/imagefactory/pull/364 15:07:09 jskladan: I'll start poking people if there's no response by later this week 15:07:22 tflink: how is the image building? I'm still mostly getting timeouts because of the cloud-init stuff 15:07:47 it's working for me and working most of the time in dev and stg 15:07:54 both should be using the image I built last week 15:08:21 * jskladan shrugs... I don't know what's wrong with me then... But as long as it works where it needs to... 15:09:04 yeah, I'm not sure what's going wrong 15:09:19 you're using the latest testcloud, right? 15:09:31 I suppose this may be better saved for not in meeting 15:09:42 any other questions, comments, infos? 15:09:48 yup, we can have a look at it once you are in Brno... 15:10:19 works for me 15:11:08 #topic Disposable Clients Release 15:11:24 I think we're mostly ready to get disposable clients into production 15:11:41 we'll have to keep a close eye on it and make sure we're manually rescheduling jobs that failed execution, though 15:12:10 the other approach would be to wait until after devconf when we're not quite so distracted 15:12:23 thoughts? 15:12:24 why do we need to manually reschedule jobs? 15:12:32 ^^ 15:13:32 because there's no way to automatically reschedule them if they fail execution 15:13:49 note that by fail execution, I mean when there's a problem with the VM or something like that 15:13:55 not if the check itself fails 15:13:59 oh, cool 15:14:18 which is probably something that should be added in the not-so-distant future :) 15:14:20 for depcheck and upgradepath, we don't mind that much, do we? 15:14:25 true 15:14:32 and rpmlint is not really important 15:14:38 good point 15:14:50 just saying, in case we want to save some time :) 15:14:52 so I'm over-thinking things :) 15:15:25 thoughts on redeploying production on Tuesday or Wednesday? 15:15:56 we also don't have a rescheduling mechanism if e.g. koji or bodhi times out. I think it's a similar situation 15:16:08 true 15:16:25 how hard is to revert if it explodes on tue/wed? 15:16:31 it'll matter more when we have more checks that aren't quite as "scheduling resilient" as upgradepath/depcheck 15:16:48 shouldn't be too bad 15:17:02 some reinstalls, git reverting of templates 15:17:29 we'll need to update all the taskotron local playbooks and docs soon, as well 15:18:02 in that case tue/wed sounds ok to me 15:18:41 ditto 15:19:43 sounds like we have a time, then 15:19:48 well, date 15:20:15 mkrizek: any thoughts on what times would work better for you? 15:20:30 assuming you're available to help with the redeploy 15:20:52 Tue works better 15:22:11 thoughts on time? 15:23:20 I am flexible, your morning/noon works for me 15:23:40 * tflink figures starting about now or doing it early-ish morning for you would make most sense 15:24:02 yep 15:24:15 any preference? 15:25:02 as a side note, I want to figure out a different way for migrating data 15:25:12 the current "tar everything up" method takes way too long 15:25:37 either works really, so let's start in the time of today's meeting? 15:25:43 sounds good to me 15:25:57 anything else on this topic? 15:26:41 not here 15:27:01 moving on, then 15:27:12 #topic Image Creation Status 15:27:17 I think this'll be pretty quick 15:28:20 Just wanted to quickly go over what will probably happen for image generation 15:28:28 My understanding is: 15:28:56 1. start with images generated manually, probably done by jskladan or me 15:29:25 2. figure out a spot on one of the qa*.qa boxes to generate images so we don't have to upload so much 15:29:43 3. get jskladan's "get newest image" script deployed to all the client hosts 15:30:06 4. get images building in an automated fashion on a daily-ish basis 15:31:41 sounds good to me 15:31:48 cool 15:31:56 that's pretty much all I had 15:32:04 which brings us to 15:32:09 #topic Open Floor 15:32:17 Any other topics which folks wanted to bring up? 15:32:43 none here 15:32:53 nope 15:33:02 not here 15:33:16 cool 15:33:25 time to get to more docs writing, etc. :) 15:33:37 * jskladan *yay* 15:34:02 * tflink is planning to do another release today to get all the docs built etc. 15:34:32 jskladan: you're excited about writing docs? 15:34:40 * tflink will have to remember that for future reference 15:34:49 :D 15:34:59 anyhow, thanks for coming everyone 15:35:04 * tflink will send out minutes shortly 15:35:15 and assign more docs tasks to jskladan :-D 15:35:19 #endmeeting